Skiml turns the pages you save into decisions, not another pile. Save any article, PDF, video, podcast, or newsletter in one click from Chrome, and Skiml writes a short action brief: a summary, the key insights, and clear action items. A daily digest brings the right saves back, and triage mode (Act, Reference, Archive) clears the backlog in minutes. Search and tags keep it all findable. Readwise resurfaces, Raindrop bookmarks, Obsidian stores; Skiml helps you finish. Free to start.
No reviews yetBe the first to leave a review for Skiml
Maker
📌
Hey Product Hunt 👋 I'm Brandon, the solo maker behind Skiml.
I've always saved way more than I read. Articles, PDFs, YouTube videos, half-finished newsletters, all piling up in a read-later app that quietly turned into a graveyard. The tools I tried were great at saving things and resurfacing them later, but none helped with the part I was actually stuck on: deciding what to do with any of it.
So I built Skiml. You save any page with one click from Chrome, and instead of a summary or a highlight you'll never reread, you get a short action brief: the key insights, a clear verdict, and what to do next. A daily digest brings the right saves back, and a keyboard-first triage mode (Act, Reference, Archive) clears the pile in minutes instead of carrying it around as guilt.
It's free to start (50 saves a month with full briefs, search, and tags), and it works across articles, PDFs, videos, podcasts, and forwarded newsletters.
I built this solo, mostly nights and weekends, and I'd genuinely love your feedback, especially the harsh kind. What would make this a daily habit for you? And if your read-later list has become a pile you feel bad about, how do you deal with it today?
Thanks for taking a look 🙏
Report
🧐 Good find
how good are the summaries actually for long technical PDFs, like do they pull out the methods section or just the abstract vibes
Report
Maker
@duranyoldze23p We pull the full PDF text server side, not just the abstract, up to about 30k characters in document order. If it's actually classified as a research paper/preprint, there's a dedicated prompt that specifically forces the brief to separate what the study found from what the authors interpret, plus call out sample size, methods, and limitations. So it's aiming for methods, not abstract vibes. Honest caveat: on a really long paper that 30k char cap can cut off before later sections, that's a real limit today but something we plan to patch later via batching for considerably longer content. That feature will be gated for pro/power users though because of the token input size.
Report
💎 Pixel perfection
Saved a long podcast and a PDF in one go, then the daily digest actually showed me the action items I'd have otherwise forgotten. Triage mode cleared my backlog in under two minutes, which never happens.
Report
Maker
@fatih1pmo This makes my day, thank you Fatih. Long podcast plus a PDF, then the digest surfacing what you'd have forgotten, that's exactly the loop I built this for. Really glad triage mode is actually clearing backlog for you.
Report
💡 Bright idea
Love how it pushes you past just saving stuff. One thing I'd want though is a way to tie action items to a specific date, like a "do this Friday" prompt, so the daily digest can flag what's actually due instead of just resurfacing older saves. Right now I can see myself still triaging instead of executing.
Report
Maker
@fratgzelsodszj Appreciate that, and that's a genuinely good idea. Right now there's no due date or schedule on action items, closest thing is marking something "act" so it stops getting resurfaced since you already own it, but there's no "do this Friday" concept yet. Adding it to the list, it'd close exactly the gap you're describing between triaging and actually executing.
Report
how does the daily digest decide what to surface first if i haven't touched anything in a few days, does it just dump everything or is there some kind of priority logic under the hood
Report
Maker
@sema1067799 Good question, there's real logic. New briefs that haven't been resurfaced yet go first, anything you've triaged to "act" never comes back (you already own it), archived saves are excluded entirely, and no single brief repeats in a digest more than once every 14 days. So it won't just dump your whole backlog on you. I'm also testing a smarter version that spreads picks across different sites/topics instead of clustering, rolling that out gradually via A/B.
Report
How does the AI decide what counts as an "action item" versus just a key insight from something like a long podcast, and does it ever get that distinction wrong in ways that need manual fixing?
Report
Maker
@meltemeskimscp It's one model call sorting everything into key insights (context/judgment stuff) vs action items (only what's concretely actionable for you, it's allowed to return zero rather than pad). For podcasts specifically it's told to treat the transcript as the real substance and tie insights to actual timestamps and who said them, not "go listen to the episode." Does it misjudge sometimes? Yeah. Right now the fix is hitting Regenerate on the whole brief, there's no way to edit just one field yet, that's rougher than I want it to be. I do plan to explore user generated brief guidelines as a future iteration though.
Report
finally something that nudges me to actually act on what i save instead of letting it rot in readwise. the act/reference/archive triage feels like a small but smart touch.
Report
Maker
@elifeuyar8v3q Thanks Elife! Yeah that was the whole idea, saving isn't the hard part, deciding what to do with it is. Really glad the triage flow is landing for you.
how good are the summaries actually for long technical PDFs, like do they pull out the methods section or just the abstract vibes
@duranyoldze23p We pull the full PDF text server side, not just the abstract, up to about 30k characters in document order. If it's actually classified as a research paper/preprint, there's a dedicated prompt that specifically forces the brief to separate what the study found from what the authors interpret, plus call out sample size, methods, and limitations. So it's aiming for methods, not abstract vibes. Honest caveat: on a really long paper that 30k char cap can cut off before later sections, that's a real limit today but something we plan to patch later via batching for considerably longer content. That feature will be gated for pro/power users though because of the token input size.
Saved a long podcast and a PDF in one go, then the daily digest actually showed me the action items I'd have otherwise forgotten. Triage mode cleared my backlog in under two minutes, which never happens.
@fatih1pmo This makes my day, thank you Fatih. Long podcast plus a PDF, then the digest surfacing what you'd have forgotten, that's exactly the loop I built this for. Really glad triage mode is actually clearing backlog for you.
Love how it pushes you past just saving stuff. One thing I'd want though is a way to tie action items to a specific date, like a "do this Friday" prompt, so the daily digest can flag what's actually due instead of just resurfacing older saves. Right now I can see myself still triaging instead of executing.
@fratgzelsodszj Appreciate that, and that's a genuinely good idea. Right now there's no due date or schedule on action items, closest thing is marking something "act" so it stops getting resurfaced since you already own it, but there's no "do this Friday" concept yet. Adding it to the list, it'd close exactly the gap you're describing between triaging and actually executing.
how does the daily digest decide what to surface first if i haven't touched anything in a few days, does it just dump everything or is there some kind of priority logic under the hood
@sema1067799 Good question, there's real logic. New briefs that haven't been resurfaced yet go first, anything you've triaged to "act" never comes back (you already own it), archived saves are excluded entirely, and no single brief repeats in a digest more than once every 14 days. So it won't just dump your whole backlog on you. I'm also testing a smarter version that spreads picks across different sites/topics instead of clustering, rolling that out gradually via A/B.
How does the AI decide what counts as an "action item" versus just a key insight from something like a long podcast, and does it ever get that distinction wrong in ways that need manual fixing?
@meltemeskimscp It's one model call sorting everything into key insights (context/judgment stuff) vs action items (only what's concretely actionable for you, it's allowed to return zero rather than pad). For podcasts specifically it's told to treat the transcript as the real substance and tie insights to actual timestamps and who said them, not "go listen to the episode." Does it misjudge sometimes? Yeah. Right now the fix is hitting Regenerate on the whole brief, there's no way to edit just one field yet, that's rougher than I want it to be. I do plan to explore user generated brief guidelines as a future iteration though.
finally something that nudges me to actually act on what i save instead of letting it rot in readwise. the act/reference/archive triage feels like a small but smart touch.
@elifeuyar8v3q Thanks Elife! Yeah that was the whole idea, saving isn't the hard part, deciding what to do with it is. Really glad the triage flow is landing for you.